Document Intake Agent for Forms, Receipts and Contracts
Plan how an agent extracts fields from documents and flags uncertainty.
Added October 15, 2026. Written with AI assistance and checked for clarity and placeholders; not reviewed by a lawyer, accountant, clinician or other professional. How these prompts are made
The Open buttons send the prompt text as currently shown below to that site only when you click. Fill in the placeholders first.
Design a document intake agent. Document types: [DOCUMENT_TYPES]. Fields to extract from each: [FIELDS]. Where documents arrive: [SOURCE]. Where results go: [DESTINATION]. Error cost (what goes wrong if a field is wrong): [ERROR_COST]. Specify: a classify step, an extract step returning JSON with value, confidence (high, medium, low) and the text snippet found, a validation step (date formats, totals that must equal the sum of lines, required fields), a human review queue for low confidence and failed validation, and a final write step with an audit entry. Provide the extraction prompt for one document type in full, with rules for missing fields (null, not a guess), multiple values, handwritten text and scanned images. Include a table of validation rules with an example failing case each. Add privacy rules: redact sensitive fields in logs, limit retention to [RETENTION], and do not send documents to tools I have not approved. Give ten test documents to prepare (clean, skewed, partly cut off, wrong type, duplicate) and the pass criteria I should set before trusting the agent.
Fill in before use
Optional: type your details and the prompt above updates. Nothing is saved or sent.
Replace each bracketed placeholder with your own details, then paste into Any AI or another assistant.